Classification using a Fuzzy Spiking Neural Network Cornelius Glackin, Liam McDaid, Liam Maguire, and Heather Sayers University of Ulster, Faculty of Engineering, School of Computing and Intelligent Systems, Magee Campus, Londonderry, BT48 7JL, Northern Ireland {glackin-cl, lj.mcdaid, lp.maguire, hm.sayers}@ulster.ac.uk Abstract This paper presents a generic spiking neural network topology that employs fuzzy reasoning using strictly biological components. The fuzzy reasoning rationale dictates the deployment of the biological components such as dynamic synapses, receptive fields and the interplay between inhibitory and excitatory neurons. Dynamic synapses are used throughout the network, enriching the flow of information and, together with receptive fields promotes neuron selectivity. The receptive fields behave like fuzzy membership functions enabling individual neurons to be selective to certain spike train frequencies. The network is supervised but learning only occurs locally as in the biological case. Learning is illustrated using the Wisconsin Breast Cancer benchmark classification problem. 1 Introduction The history of neural network research is characterised by a progressively greater emphasis paid to biological plausibility. The evolution of neuron modelling with regard to the complexity of computational units can be classified into three distinct generations [1]. The third generation of neuron modelling (spiking neurons) is based on the realisation that the precise mechanism by which biological neurons encode and process information is poorly understood. The spatio-temporal distribution of spikes in biological neurons ‘holds the key’ to understanding the brain’s neural code. There exists a multitude of spiking neuron models that can be employed in spiking neural networks (SNNs). The models range from the computationally efficient on the one hand to the biologically accurate on the other [2]. The extensive amount and variety of neuron models [2] exist in acknowledgement of the fact that there is a trade-off between the individual complexity of spiking neurons and the number of neurons that can be modelled in a neural network. In addition to the variety of neuron models, biological neurons can have two different roles to play in the flow of information within neural circuits. These two roles are excitatory and inhibitory respectively. Excitatory neurons are responsible for relaying information whereas inhibitory neurons locally regulate the activity of excitatory neurons. Ongoing physiological experiments continue to illuminate the underlying processes responsible for the complex dynamics of biological neurons. The degree to which these complex dynamics are modelled in turn limits the size and computational power of SNNs. Therefore it is imperative to determine which biological features improve computational capability whilst enabling an efficient description of neuron dynamics. Ultimately neuro-computing seeks to implement learning in a human fashion. In any kind of algorithm where human expertise is implicit, fuzzy IF-THEN rules provide a language for describing this expertise [3]. In this paper, the rationale for the distribution of biologically-inspired computational elements is prescribed by the implementation of fuzzy IF-THEN rules. In Section 2, unsupervised and supervised learning methods, dynamic synapses and receptive fields are reviewed. Section 3 includes a brief discussion of how fuzzy reasoning can provide a basis for structuring the network topology and introduces a generic network topology outlining the specific models and algorithms used to implement fuzzy reasoning. Experimental results and remarks for the Wisconsin Breast Cancer classification problem are given in Section 4, and conclusions and future research directions are presented in Section 5. 2 Review The modelling of the neuron’s synapse is an essential aspect for an accurate representation of real neurons, and one of the key mechanisms to reproducing the plethora of neuro-computational artefacts in SNNs. From a biologically plausible point-of-view synaptic modification in spiking neurons should be based on the temporal relationship between pre and post-synaptic neurons, in accordance with Hebbian principles [4]. In fact, Hebbian learning Proceedings of the 2008 UK Workshop on Computational Intelligence Page 117